#rest on irc.freenode.net 2012-01-03 (txt)

[01:57] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) joined #rest.
[01:58] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) left #rest.
[02:30] Wombert (~Wombert@client-gate.sfa.network.cynigram.com) joined #rest.
[02:38] Wombert (~Wombert@client-gate.sfa.network.cynigram.com) left irc: Quit: Wombert
[02:57] mr_yall (~mr_yall@scifo.mexico86.co.uk) joined #rest.
[04:15] <mikekelly> where the feck is mamund
[04:16] <mikekelly> please can someone talk me through the point that Jan is making here because I really cannot fathom it :(
[04:17] <mikekelly> http://tech.groups.yahoo.com/group/rest-discuss/message/18325
[04:56] ssedano (~ssedano@unaffiliated/ssedano) left irc: Ping timeout: 240 seconds
[05:10] ssedano (~ssedano@unaffiliated/ssedano) joined #rest.
[05:18] <mikekelly> have we had a rest-discsuss thread recently that hasn't been hijacked by people seemingly desperate to tell other people they don't know what they are talking about?
[05:21] <mikekelly> I think a big problem is that many people in the community (regardless of whether they are wrong or right) just do not express themselves in a way that average lay-people can comprehend
[05:24] <mikekelly> in my experience, people who express themselves in complicated ways are usually compensating for mental models that aren't cohesive or coherent.
[05:39] <mogsie> mikekelly: Could it be terminology differences?
[05:40] <mikekelly> well I did spot his 'server' vs 'resource' processing comments
[05:40] <mogsie> e.g. "Application" to Jan (and Fielding) means e.g. "buying a book at Amazon" not "MicrosoftWord" or "Internet Explorer"
[05:40] <mikekelly> no no that's my interp of application
[05:40] <mogsie> The Application is therefore the result of interacting with a web site through a browser.
[05:40] <mogsie> That's what I expected :-)
[05:40] <mikekelly> I think Jan's interp is Application means layer 7 on OSI
[05:41] <mogsie> That may be right.
[05:41] <mikekelly> but I did point out that 'hypertext as the engine of applicaiton state' doesn't really fit into that interp
[05:41] <mogsie> So when you say Application, Jan might think OSI7 Application...
[05:41] <mikekelly> i.e. there are lots of different uses of 'application' depending on the context
[05:42] <mogsie> When you click "Buy this book" link on amazon, you're driving the application forward.
[05:42] <mogsie> The application being "buying a book"...
[05:42] <mogsie> What is your interp of Application?
[05:54] <mikekelly> "a set of transitions a client completes to achieve an objective"
[05:54] <mikekelly> something like that
[05:54] <mikekelly> in this context it doesn't have anything to do with Layer 7, put it that way.
[06:01] <mogsie> Well, I'm not sure if I understand what you mean by an "OSI layer 7" interpretation of the term appllication...
[06:02] <mogsie> How is my definition (result of interacting with a web site through a browser) different than yours (a set of transitions that a client completes to achieve an objective)?
[06:15] scudco (~scudco@cpe-75-85-13-152.socal.res.rr.com) joined #rest.
[06:46] <mikekelly> mogsie: it's not, I'm agreeing with you
[06:46] <mikekelly> the important point there for me is that it is about *you* and not your browser
[06:46] <mikekelly> that is actually something people get horribly confused about
[06:48] <mikekelly> css,javascript,etc - all the hypermedia stuff your browser does are generic intermediary mechanisms that produce the generic interface to us (people) that we call a 'web page'
[06:49] <mikekelly> they are intermediary mechanisms - and that's why the hypertext in them is standardised and the semantics contained in the HTML spec
[06:50] <mikekelly> so my point in all of this discussion is; if you can find a generic media type that provides the hypertext controls to express your application and facilitates the intermediary mechanisms you need
[06:50] <mikekelly> don't bother minting a new media type, it's waste of effort.
[06:51] <mikekelly> hence why when you expose a web application you don't create a new media type or extend html.
[06:52] <mikekelly> I don't really see what is controversial about that point of view - it reflects reality.
[06:54] <mikekelly> what *really* pisses me off about this discussion is that Jan keeps referring to 'visibility'
[06:55] <mikekelly> and then when I press him on what exactly he means by that and why it presents and issue
[06:55] <mikekelly> he just ignores me.
[06:57] <mikekelly> possibly because 'visibility' is actually being used there a meaningless buzzword, to try and imply there's a loss of something valuable when there isn't.
[06:57] <mikekelly> which is another thing I have asked about, and had no response to.
[06:57] <fu-manchu> I admit I don't get Jan's point of view in this recent discussion either
[06:57] <fu-manchu> "When was the last time you implemented a POST handler of a resource and thought in any way about what the client expects?"
[06:57] <fu-manchu> my answer is "always"
[06:58] <mikekelly> I don't mind people dissagreeing with me provided they aren't being so fucking unnecessarily condescending.
[06:58] <fu-manchu> not "never"
[06:58] <mikekelly> this is why I accused him of pedantry
[06:58] <mikekelly> I think there's some stupid Zen-master shit going on there
[06:59] <mikekelly> pseudo Zen-master*
[06:59] <mikekelly> :)
[06:59] <trygvis> fu-manchu: really? does as in that you care about the current state of the client or what?
[06:59] <mikekelly> well what's a functional test if it's not some expectation ?
[07:00] <trygvis> to me his point seems perfectly valid, the server only care about the message that was POSTed, not the internal state of the client
[07:00] <mikekelly> that's not what the question says :|
[07:00] <fu-manchu> maybe I need to read that email again; I didn't see anything about "internal state" in there
[07:00] <mikekelly> there isn't, I've read almost all his posts several times
[07:01] <trygvis> it might be what I read into the message
[07:02] <mikekelly> It seems like there's a lot of word-trickery and pedantry going on which confuses everything a lot
[07:02] <mikekelly> am I the only person that gets pissed off with that?! :(
[07:03] <mikekelly> I find it such a frustrating waste of time
[07:05] <fu-manchu> I wonder if he imagines (or uses daily) some framework where "the server" is a separate component from "each resource", where "the server" component is little more than a deserializer of the HTTP request
[07:06] <fu-manchu> and a "POST handler" therefore hands off an internal object representation to the *real* resource component
[07:07] <mikekelly> yeah - an intermediary :|
[07:08] <mikekelly> that being the case, I think that model is just horribly confusing and adds nothing
[07:08] <mikekelly> it's definitely not a standard way of talking about a 'server'
[07:08] <mikekelly> when I say server I mean the-thing-that-takes-the-request-and-produces-a-response
[07:08] <fu-manchu> I might be putting words in his mouth
[07:08] <mikekelly> no no that's exactly where I thought he was going
[07:09] <mikekelly> so it makes 2 of us independently
[07:09] <mikekelly> I have actually asked him explicitly about it so I might get a response
[07:10] <mikekelly> I would really hate to be someone relatively new to REST and reading this bloody thread.
[07:11] <mikekelly> nasty business
[07:35] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Ping timeout: 244 seconds
[07:50] mamund (mamund@nullrouted.info) joined #rest.
[07:50] #rest: mode change '+o mamund' by ChanServ!ChanServ@services.
[07:51] mamund is here!
[07:59] mamund is drawing mikekelly into his net<g>
[07:59] <mikekelly> ¬_¬
[07:59] <mikekelly> I'm game, let's do this.
[07:59] <mamund> LOL
[08:00] <mamund> you've already done it<g>
[08:06] <mamund> and the trap door slowly closes....
[08:08] <mikekelly> it's not really a trap if I walk in and help you close it.
[08:08] <mikekelly> ;)
[08:08] <mamund> :)
[08:08] <mikekelly> you're just trying to hijack the thread
[08:08] <mamund> true
[08:08] <mikekelly> admit it!
[08:08] <mikekelly> ;)
[08:08] <mikekelly> oh you did.
[08:08] <mamund> that's all i needed
[08:09] <mamund> srsly, you are building hypermedia for machines. that's all i'm talking about
[08:10] <mikekelly> yeah my post was about the realistic'ness of trying to emulate human-like behaviour with automated clients
[08:10] <mikekelly> right ?
[08:11] <mikekelly> maybe it wasn't clear in the post and we should clarify on the list but basically I (on purpose) focused on what I would term the 'realistic' hypermedia control as far as machines are concerned
[08:11] <mikekelly> at this point in time
[08:12] <mikekelly> maybe in the future there will be some development in programming paradigms that dramatically improves the cost of developing more sophisticated clients
[08:13] <mikekelly> until that happens I think exposing a distributed application with the assumption that has already happened
[08:13] <mikekelly> is like pissing into the wind
[08:13] <mikekelly> i.e. not a good idea and will leave you in a mess
[08:14] <mikekelly> either assuming too much about your clients or raising the barrier so high they never actually become your client.
[08:14] <mamund> yeah, i agree that it's not realistic to assume that machine-based hypermedia will have the same features/functionality of human-based hypermedia
[08:14] aGHz (~Adium@modemcable028.248-23-96.mc.videotron.ca) joined #rest.
[08:14] <mamund> i think we've covered this territory in IRC before<g>
[08:14] <mikekelly> indeed.
[08:15] <mikekelly> let's do it all over again on the list anyway, might as well.
[08:15] <mikekelly> ¬_¬
[08:15] <mamund> my point (in the mailing list) is simply that we should be able to ack. that machine-based hypermedia is not only possible, but with us already.
[08:15] <mamund> we should also recognize that it is "not the same" as human-based hypermedia
[08:15] <mikekelly> we can make it short and I can ask you for some evidence of 'dramatically improved evolvability' from form mechanisms (i.e. more than just aesthetically altered output from clients)
[08:16] <mamund> LOL
[08:16] <mamund> sure
[08:16] <mikekelly> be honest, it's not there.
[08:16] <mamund> no, it's not
[08:16] <mamund> and, IMO, doesn't _need_ to be there
[08:16] <mikekelly> PRAISE THE LAWWWWWD!
[08:16] <mamund> sh*t son, you've heard that from me for more than a year
[08:17] <mikekelly> :)
[08:17] <mamund> :)
[08:17] <mikekelly> yeah I guess my post in response first skipped forward a few posts that never made it into existence
[08:18] mikekelly shrugs
[08:18] <mamund> heee
[08:18] <mikekelly> I do that all the time
[08:18] <mamund> multi-tasking over time!
[08:18] <mikekelly> no, I'm just a moron
[08:18] <mikekelly> did you spot by RDF dig ?
[08:18] <mikekelly> I really should've started a new thread for that :)
[08:19] <mamund> :)
[08:19] <mikekelly> s/by/my
[08:20] <mikekelly> why does Jan keep using the terminology "This is where you get it wrong" ?
[08:20] <mikekelly> who does that? :|
[08:20] <mamund> no clue
[08:21] <mamund> he's kinda got himself in a hole here
[08:21] <mamund> bummer
[08:21] <mamund> that thread is wandering between describing protocol-level semantics and app-level semantics...
[08:21] <mikekelly> it's difficult to know what to do about this thread because it does seem like there are some interesting underlying bits but it just keeps veering off into inane semantic arguments
[08:21] <mamund> not interested in helping sort it out, either.
[08:22] <mikekelly> I'm honestly trying to avoid doing that
[08:22] <mikekelly> it's the word 'application'
[08:22] <mamund> yeah, i am not participating in that "part" of the thread<g>
[08:23] <mikekelly> seems like the get-out clause for certain parties in that thread is to pretend like people are talking about layer 7 when they say "application"
[08:23] <mamund> again, all a mess; not interestd
[08:23] <mikekelly> lol fair enough
[08:23] <mikekelly> I'll shut up now
[08:23] <mikekelly> :)
[08:23] <mamund> ha!
[08:24] <mikekelly> not on the mailing list
[08:24] <mikekelly> just here
[08:24] <mamund> Heeeeeee
[08:24] <mikekelly> who do you think I am ?!
[08:24] <mikekelly> ok ok mamund
[08:24] <mikekelly> please
[08:24] <mikekelly> one quote and then no more
[08:25] <mikekelly> please!!
[08:25] <mikekelly> :D
[08:25] <mamund> :)
[08:25] <mamund> lay it on me, dude
[08:25] <mikekelly> > - What the resource does upon POST is determined by understanding the
[08:25] <mikekelly> > semantics of the application it is part of
[08:25] <mikekelly> No, this is where you get it wrong. The resource has no idea what the application is. Especially is it not bound to whatever the client expects.
[08:25] <mikekelly> quote is me, last line is Jan
[08:25] <mamund> again, the context-switching here, ick
[08:26] <mikekelly> that's one crazy-ass context he's switching into
[08:26] <mamund> honestly, even in these lines, it's all a mess.
[08:26] <mamund> first, "the resource _does_" ?!~@
[08:26] <mamund> resources don't do aquat, eh?
[08:26] <mamund> a/aquat/squat
[08:27] <mikekelly> right I'd have used server
[08:27] <mikekelly> he brought in the resource
[08:27] <mikekelly> apparently don't do anything
[08:27] <mikekelly> resources to stuff
[08:27] <mamund> POST is protocol, Jan is talking application, eww
[08:27] <mikekelly> look, I don't know - I'm just trying to strip away the bullshit and have a proper conversation
[08:27] <mikekelly> :D
[08:27] <mikekelly> er
[08:27] <mikekelly> well
[08:27] <mikekelly> POST is part of the 'Application' in the OSI sense
[08:28] <mamund> in my mind there are protocol-level "semantic understandings"
[08:28] <mikekelly> yeah I agree with you, btw I just said "yeah" out loud
[08:28] <mamund> and there are message-level "semantic understandings"
[08:29] <mamund> i see that
[08:29] <mikekelly> protocol stuff is the stuff of 'visibility' right ?
[08:29] <mamund> yes
[08:29] <mikekelly> ok good I'm not going insane.
[08:29] <mamund> what happens in the message, stays in the message<g>
[08:29] <mikekelly> indeed.
[08:30] <mikekelly> ah man I feel infinitely better already.
[08:30] <mamund> no, you're not insane (at least not due to this silly thread)
[08:30] <mamund> !
[08:30] <mikekelly> :D
[08:47] <mamund> and closing out my thread-jacking for the day
[08:58] <mikekelly> ha
[08:59] <mikekelly> incredibly it looks like someone is agreeing with what I'm saying
[08:59] <mikekelly> extraordinary.
[09:06] KevBurnsJr (~KevBurnsJ@50.0.103.39) joined #rest.
[09:12] <mamund> :)
[09:34] whartung (~whartung@wsip-98-189-78-118.oc.oc.cox.net) joined #rest.
[09:34] <whartung> HNY all
[09:34] <mamund> o/
[09:37] <mikekelly> and to you
[10:07] technoweenie (~technowee@host-86-220-9-69.midco.net) joined #rest.
[10:51] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #rest.
[10:56] <mamund> ha! two (tiny) commits to Cloud9 in the last week. i haz OSS skilz!
[10:57] mamund struts comically
[10:59] <mikekelly> mamund: heh, I looked at cloud9 but I really need it to have better vim bindings
[10:59] <mikekelly> when I tried it last, the vim stuff was kind of poop
[10:59] <mamund> vim stuff is still weak sauce
[11:00] <mamund> but it has gotten better in the last month
[11:00] <mikekelly> cloud9 is the one where you import 'projects' from github and work on them direct right ?
[11:00] <mamund> IIRC, no ex commands at all
[11:22] <mamund> mikekelly: actually just played around w/ latest vim bindings for C9; seem much improved
[11:23] <mamund> since i'm still a VIm n00b, would be curious to hear what you think of it right now.
[11:40] ex_nihilo (ex_nihilo@gateway/shell/blinkenshell.org/x-rpfetmzovhvollns) joined #rest.
[11:42] dkubb (~dkubb@50.92.131.237) joined #rest.
[12:01] dkubb (~dkubb@50.92.131.237) left irc: Quit: Linkinus - http://linkinus.com
[12:54] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Computer has gone to sleep.
[13:05] <mikekelly> mamund: hmm it's ok - the basics work ok
[13:06] <mikekelly> there's no visual block mode
[13:09] <mikekelly> If I go into visual mode and move to a line number it doesn't create a selection
[13:09] <mikekelly> it is very cool but I'm too used to all the snazzy stuff I get from vim
[13:10] <mamund> yeah, block mode. that's right
[13:10] <mikekelly> to be fair I don't use visual block mode that much
[13:10] <mikekelly> it was just one of the things I thought to test :)
[13:10] <mamund> you can still highlight w/ arrow keys, etc.
[13:10] <mamund> yep
[13:11] <mikekelly> yeah visual block mode is the visual mode where you can edit multiple lines at once
[13:11] <mikekelly> atm I use my laptop for dev work
[13:11] <mikekelly> I was thinking about moving my entire dev setup onto a box and just ssh'ing onto that
[13:13] <mikekelly> can't make my mind up, I really like the nice lookingness of MacVim which I'll lose if I run vim in a terminal
[13:13] <mikekelly> such a tough life.
[13:13] <mamund> been vim-terminalling for a cuople weeks (forcing myself, actually) and getting used to it<g>
[13:13] <whartung> Theres's vim outside of a terminal??
[13:13] <whartung> Why?
[13:14] <mamund> :)
[13:20] <mikekelly> native clipboard access
[13:20] <whartung> I can copy/paste using the terminal.
[13:20] <whartung> but I just scrape it
[13:20] <mikekelly> another modifier key w/ apple key
[13:20] <mikekelly> nice looking fonts
[13:21] <whartung> if I do "10dd" I don't get that yanked in to the native clipboard.
[13:23] <mikekelly> technoweenie: are you a rails person ?
[13:24] <technoweenie> i am
[13:24] <technoweenie> well, i was. i dont work on the framework any more
[13:24] <mikekelly> what do you think about something like this https://gist.github.com/1547058
[13:24] <mikekelly> basically I renamed the resources keyword to collection in the routing dsl
[13:25] <technoweenie> there are plugins that do similar stuff. i personally dont like them though
[13:25] <mikekelly> and created a collection controller that has collection and member blocks in it
[13:25] <mikekelly> technoweenie: what is it you don't like about that?
[13:26] Hakon|mbp (~hakon1@193.137.202.84.customer.cdi.no) joined #rest.
[13:33] Hakon|mbp (~hakon1@193.137.202.84.customer.cdi.no) left irc: Ping timeout: 268 seconds
[13:48] <technoweenie> i dont like stacking a bunch of magic framework stuff on it like that
[13:48] <technoweenie> some people do, more power to them.
[13:48] <technoweenie> i'd suggest making it a plugin/engine thing
[15:01] mamund enough cage-rattling for the day. g'nite all!
[15:11] hobodave (~hobodave@pdpc/supporter/professional/hobodave) left irc: Quit: Textual IRC Client: http://www.textualapp.com/
[15:25] technoweenie (~technowee@host-86-220-9-69.midco.net) left irc: Remote host closed the connection
[15:33] kennethreitz (~kennethre@static-71-166-241-209.washdc.east.verizon.net) joined #rest.
[16:03] kennethreitz (~kennethre@static-71-166-241-209.washdc.east.verizon.net) left irc: Quit: Computer has gone to sleep.
[16:17] KevBurnsJr (~KevBurnsJ@50.0.103.39) left irc:
[16:37] prnightmare (~will@dhip-088.east-quad.residences.colby.edu) joined #rest.
[17:26] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[17:56] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Remote host closed the connection
[18:02] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[18:08] aGHz (~Adium@modemcable028.248-23-96.mc.videotron.ca) left irc: Quit: Leaving.
[18:09] sbanwart_ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[18:09] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Ping timeout: 268 seconds
[18:16] sbanwart_ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Remote host closed the connection
[18:16] sbanwart_ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[18:36] prnightmare (~will@dhip-088.east-quad.residences.colby.edu) left #rest.
[19:02] postmodern (~postmoder@c-71-237-178-98.hsd1.or.comcast.net) joined #rest.
[19:03] <postmodern> has any RESTful web framework, allowed you to query the URL template
[19:03] <postmodern> and get back descriptions for the params in the URL?
[19:09] <mamund> postmodern: hey there, not sure of your Q...
[19:09] <mamund> are you asking if any web frameworks implement the URI Template Internet Draft?
[19:09] <mamund> or is this a general "any templated URI" frameworks Q?
[19:09] <postmodern> mamund, well if any of them have a way to return the possible URI paths it supports
[19:10] <postmodern> mamund, along with descriptions of each one
[19:10] <mamund> ahh
[19:10] <mamund> well, not that i know of.
[19:10] <mamund> you might want to check the implementing-rest project on google code
[19:10] <mamund> there are a number of framework listed there
[19:10] <postmodern> "/books/:isbn" => {'isbn': {type: "Integer", desc: "ISBN of the book"}}
[19:11] <mamund> in waht way do you want to use this feature?
[19:12] <mamund> of source, HTML does this, right?
[19:12] <mamund> HTML.FORM@method="get" is a URI template w/ descriptors inline
[19:12] <postmodern> mamund, i want to provide a way for a client to introspect the RESTful routes
[19:12] <mamund> hmmm
[19:12] <mamund> you mean a human or a machine?
[19:12] <postmodern> probably both, since i included a description
[19:13] <mamund> well, shouldn't be too hard to whip up in any framework
[19:14] <mamund> you want some list of the currently available queries, then?
[19:14] <postmodern> yeah
[19:14] <mamund> where "currently-avail" === "at this point in time for this current user/session"
[19:14] <postmodern> and if other frameworks have implemented this, and how
[19:15] <mamund> yeah, best i can come up w/ now is HTML.FORMs
[19:15] <mamund> you could add a link to any response (either inline or link header) that points to some
[19:15] <mamund> document, etc. on the server that will return that info.
[19:15] <mamund> assuming your server "knows" that info<g>
[19:16] aGHz (~Adium@modemcable153.0-178-173.mc.videotron.ca) joined #rest.
[19:16] <mamund> or just include it in the response; why have an external link?
[19:17] <postmodern> yeah i was just asking if others have done this, and how did they allow querying that information
[19:17] <mamund> yeah
[19:17] <mamund> i did a custom media type design in JSON that has query templates.
[19:18] <mamund> they can be inserted into the response or a single link can appear so the client can pull the templates that way
[19:19] <mamund> you can check it out here: http://amundsen.com/media-types/collection/
[19:19] <mamund> might give you some ideas
[20:50] sbanwart_ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Ping timeout: 255 seconds
[22:54] ex_nihilo (ex_nihilo@gateway/shell/blinkenshell.org/x-rpfetmzovhvollns) left irc: Quit: Lost terminal
[23:02] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #rest.
[00:00] --- Wed Jan 4 2012